> ## Documentation Index > Fetch the complete documentation index at: https://apixo.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # InfiniteTalk API > Audio-driven talking avatar video generation from a portrait and audio ## Overview InfiniteTalk turns a source portrait plus an audio track into a talking or singing avatar video with lip synchronization. Use this page when you are ready to call the API after trying the model in the APIXO playground. | Capability | Value | | ------------- | ---------------------------------------------------- | | Model ID | `infinitetalk` | | Mode | `image-to-video` | | Source images | 1 required image URL, plus 1 optional mask image URL | | Audio input | Exactly 1 public HTTP(S) MP3, WAV, or M4A URL | | Audio limits | Up to 128 MB and 600 seconds | | Resolutions | `480p`, `720p` | | Billing | Per probed audio second, minimum 5 seconds | ## Endpoint and authentication Base URL: ```text theme={null} https://api.apixo.ai/api/v1 ``` | Method | Endpoint | Purpose | | ------ | ------------------------------------------ | ------------------------------------- | | `POST` | `/generateTask/infinitetalk` | Submit a generation task | | `GET` | `/statusTask/infinitetalk?taskId={taskId}` | Poll task status and retrieve results | All requests require your APIXO API key: ```http theme={null} Authorization: Bearer YOUR_API_KEY ``` Submit requests also require: ```http theme={null} Content-Type: application/json ``` ## Copy-paste async quickstart This minimal request submits an InfiniteTalk task and returns a `taskId`. ```bash theme={null} curl -X POST "https://api.apixo.ai/api/v1/generateTask/infinitetalk" \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "request_type": "async", "input": { "mode": "image-to-video", "image_urls": [ "https://example.com/portrait.png" ], "audio_urls": [ "https://example.com/voice.mp3" ], "prompt": "make the portrait speak naturally with subtle head motion", "resolution": "480p", "seed": -1 } }' ``` Successful response: ```json theme={null} { "code": 200, "message": "success", "data": { "taskId": "task_12345678" } } ``` Save the `taskId`; you need it to poll for the final result. ## Poll for result ```bash theme={null} curl -X GET "https://api.apixo.ai/api/v1/statusTask/infinitetalk?taskId=task_12345678" \ -H "Authorization: Bearer YOUR_API_KEY" ``` Processing response: ```json theme={null} { "code": 200, "message": "success", "data": { "taskId": "task_12345678", "state": "processing", "createTime": 1767965610929 } } ``` Success response: ```json theme={null} { "code": 200, "message": "success", "data": { "taskId": "task_12345678", "state": "success", "resultJson": "{\"resultUrls\":[\"https://file.apixo.ai/video.mp4\"]}", "createTime": 1767965610929, "completeTime": 1767965652317, "costTime": 41388 } } ``` Failed response: ```json theme={null} { "code": 200, "message": "success", "data": { "taskId": "task_12345678", "state": "failed", "failCode": "SensitiveContent", "failMsg": "Content violates safety policy, please adjust the prompt", "createTime": 1767965610929, "completeTime": 1767965620132 } } ``` Parse `resultJson` after `state` becomes `success`: ```javascript theme={null} const payload = JSON.parse(data.resultJson); const videoUrls = payload.resultUrls; ``` ## Request body ### Image-to-video ```json theme={null} { "request_type": "async", "input": { "mode": "image-to-video", "image_urls": [ "https://example.com/portrait.png" ], "audio_urls": [ "https://example.com/voice.m4a" ], "prompt": "keep the face stable while syncing lips to the audio", "resolution": "480p", "seed": -1 } } ``` ### With an optional mask image ```json theme={null} { "request_type": "async", "input": { "mode": "image-to-video", "image_urls": [ "https://example.com/portrait.png", "https://example.com/mask.png" ], "audio_urls": [ "https://example.com/voice.wav" ], "prompt": "subtle head motion, natural eye movement", "resolution": "720p", "seed": 7 } } ``` `image_urls[0]` is sent as the source image. `image_urls[1]`, when provided, is sent as the mask image. ## Parameters Result delivery mode. Use `async` for polling with `statusTask`, or `callback` for webhook delivery. Required when `request_type` is `callback`. Use a public endpoint that can receive the final task payload. See [Webhooks](/docs/api-reference/webhooks). InfiniteTalk input parameters. Generation mode. Only `image-to-video` is supported. Image URLs. Supports 1 or 2 URLs. The first URL is the source portrait; the optional second URL is forwarded as `mask_image`. Audio URL list. Must contain exactly 1 public HTTP(S) URL. The audio must be MP3, WAV, or M4A, up to 128 MB, and no longer than 600 seconds. Optional prompt for expression, motion, or scene guidance. If provided, it must be a non-empty string. Output resolution. Supported values: `480p`, `720p`. Values are normalized to lowercase. Random seed. Use `-1` for random output, or an integer from `0` to `2147483647` for reproducible generation. ## Response format ### Submit task response `POST /generateTask/infinitetalk` returns a task ID when the task is accepted: API status code. `200` means the task was accepted. Human-readable status message. Unique task identifier used with the status endpoint. ### Status response fields Unique task identifier. Current task state: `pending`, `processing`, `success`, or `failed`. JSON string containing the generated video URLs as `resultUrls`. Present when `state` is `success`. Machine-readable failure code. Present when `state` is `failed`. Human-readable failure message. Present when `state` is `failed`. Task creation timestamp in Unix milliseconds. Task completion timestamp in Unix milliseconds. Present after completion. Processing time in milliseconds. Present after successful completion when available. ## Webhook callback mode Use callback mode when your backend should receive the final result automatically instead of polling. ```bash theme={null} curl -X POST "https://api.apixo.ai/api/v1/generateTask/infinitetalk" \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "request_type": "callback", "callback_url": "https://your-server.com/webhooks/apixo", "input": { "mode": "image-to-video", "image_urls": [ "https://example.com/portrait.png" ], "audio_urls": [ "https://example.com/voice.mp3" ], "prompt": "natural speech with subtle head movement", "resolution": "720p", "seed": -1 } }' ``` Callback payloads use the same `code`, `message`, and `data` shape as the status response. See [Webhooks](/docs/api-reference/webhooks) for delivery requirements and retry behavior. ## Billing InfiniteTalk is billed per probed audio second. The backend downloads the audio, probes its real audio length, then bills: ```text theme={null} billableSeconds = max(5, actualAudioSeconds) totalCost = billableSeconds * resolutionUnitPrice ``` | Resolution | APIXO price | | ---------- | ---------------- | | `480p` | `$0.03 / second` | | `720p` | `$0.06 / second` | For current route and market comparison pricing, see [Pricing](https://apixo.ai/pricing). ## Latency and polling Actual latency varies by audio length, selected resolution, upstream queue load, and final result storage. Longer audio generally takes longer to complete. | Audio length | Recommended first poll | Poll interval | | ------------ | ------------------------ | ------------- | | Up to 30s | 30s after task creation | 10s-15s | | 30s-120s | 60s after task creation | 15s-30s | | 120s-600s | 120s after task creation | 30s | For production workloads, especially long audio, use callback mode to avoid frequent polling. Rate limits and concurrency can vary by account, API key, and route. If you receive `429`, slow down requests and retry with backoff. For account-level details, see [System APIs](/docs/api-reference/system). ## Errors and troubleshooting ### HTTP errors | Code | Meaning | What to do | | ----- | ---------------------------------------------------------------------------------------------------------------- | ------------------------------------------------ | | `400` | Invalid request body, unsupported mode, invalid parameter type or value, unreachable URL, or audio probe failure | Fix the request before retrying | | `401` | Missing or invalid API key | Check the `Authorization` header | | `402` | Insufficient balance or quota | Add balance or switch account/key | | `403` | Key or route cannot access the model | Check permissions and route strategy | | `429` | Rate limit or concurrency limit reached | Retry with exponential backoff | | `500` | Server error or unmapped upstream failure | Retry with backoff | | `502` | Upstream service error | Retry with backoff | | `504` | Upstream timeout | Retry or use callback mode for long-running jobs | ### Common validation issues | Issue | What happened | What to do | | ------------------------------ | ---------------------------------------------------- | ------------------------------------------------------- | | Missing `input.mode` | InfiniteTalk requires `image-to-video` mode | Set `input.mode` to `image-to-video` | | Invalid `image_urls` length | The backend accepts 1 or 2 image URLs | Send the portrait URL first, then the optional mask URL | | Invalid `audio_urls` length | The backend accepts exactly 1 audio URL | Send a single MP3, WAV, or M4A URL | | Audio cannot be fetched | The backend must fetch the audio before billing | Use a public HTTP(S) direct URL | | Audio is too large or too long | Audio is limited to 128 MB and 600 seconds | Shorten or compress the audio | | Empty prompt | `prompt` is optional, but cannot be blank if present | Remove the field or send a non-empty string | | Invalid seed | `seed` must be an integer from `-1` to `2147483647` | Use `-1` or a valid integer | ### Task failure codes | Fail code | Meaning | What to do | | ---------------------------------- | -------------------------------------------------------- | ------------------------------------------------------- | | `SensitiveContent` | Upstream rejected the prompt or media for safety reasons | Change the prompt, source image, or audio | | `PromptInvalid` | Upstream rejected or could not use the prompt | Simplify the prompt and retry | | `RateLimited` | Upstream rate limited the task | Retry with backoff | | `Timeout` | Upstream timed out | Retry, reduce input complexity, or use callback mode | | `MissingParameter` or `BadRequest` | Upstream reported a malformed request | Recheck the request body | | `Unknown error` | Failure could not be mapped to a known code | Retry with backoff or contact support with the `taskId` | See [Error Codes](/docs/api-reference/errors) for the full error reference. ## Related links * [Generation API Overview](/docs/models) * [Video Models](/docs/models/video) * [Generate Task](/docs/api-reference/generate-task) * [Status Task](/docs/api-reference/status-task) * [Webhooks](/docs/api-reference/webhooks) * [Error Codes](/docs/api-reference/errors) * [Parameter Specification](/docs/api-reference/parameters) * [Routing Strategies](/docs/concepts/routing-strategies) * [Try InfiniteTalk in the APIXO Playground](https://apixo.ai/models/infinitetalk) * [Pricing](https://apixo.ai/pricing)